LEEP: A New Measure to Evaluate Transferability of Learned Representations

Abstract

We introduce a new measure to evaluate the transferability of representations learned by classifiers. Our measure, the Log Expected Empirical Prediction (LEEP), is simple and easy to compute: when given a classifier trained on a source data set, it only requires running the target data set through this classifier once. We analyze the properties of LEEP theoretically and demonstrate its effectiveness empirically. Our analysis shows that LEEP can predict the performance and convergence speed of both transfer and meta-transfer learning methods, even for small or imbalanced data. Moreover, LEEP outperforms recently proposed transferability measures such as negative conditional entropy and H scores. Notably, when transferring from ImageNet to CIFAR100, LEEP can achieve up to 30% improvement compared to the best competing method in terms of the correlations with actual transfer accuracy.

arXiv commentPublished at the International Conference on Machine Learning (ICML) 2020

References in corpus (14)

• MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications
• Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithms
• YOLOv3: An Incremental Improvement
• DeCAF: A Deep Convolutional Activation Feature for Generic Visual Recognition
• How transferable are features in deep neural networks?
• MXNet: A Flexible and Efficient Machine Learning Library for Heterogeneous Distributed Systems
• Continual Learning Through Synaptic Intelligence
• Domain Adaptation: Learning Bounds and Algorithms
• GluonCV and GluonNLP: Deep Learning in Computer Vision and Natural Language Processing
• LEEP: A New Measure to Evaluate Transferability of Learned Representations
• Regularized Learning for Domain Adaptation under Label Shifts
• FixyNN: Efficient Hardware for Mobile Computer Vision via Transfer Learning
• Toward Understanding Catastrophic Forgetting in Continual Learning
• Improving and Understanding Variational Continual Learning

Cited by in corpus (20)

• LEEP: A New Measure to Evaluate Transferability of Learned Representations
• LogME: Practical Assessment of Pre-trained Models for Transfer Learning
• Frustratingly Easy Transferability Estimation
• Newer is not always better: Rethinking transferability metrics, their peculiarities, stability and performance
• Ranking and Tuning Pre-trained Models: A New Paradigm for Exploiting Model Hubs
• A linearized framework and a new benchmark for model selection for fine-tuning
• Aggregate or Not? Exploring Where to Privatize in DNN Based Federated Learning Under Different Non-IID Scenes
• How to Estimate Model Transferability of Pre-Trained Speech Models?
• A Multi-Stage Attentive Transfer Learning Framework for Improving COVID-19 Diagnosis
• CrowdTransfer: Enabling Crowd Knowledge Transfer in AIoT Community
• Scalable Multitask Learning Using Gradient-based Estimation of Task Affinity
• A Channel Coding Benchmark for Meta-Learning
• Similarity of Classification Tasks
• Ranking Neural Checkpoints
• A partition-based similarity for classification distributions
• Practical Transferability Estimation for Image Classification Tasks
• Transferability Estimation for Semantic Segmentation Task
• A Multiplexed Network for End-to-End, Multilingual OCR
• Exploiting a Zoo of Checkpoints for Unseen Tasks
• Probabilistic task modelling for meta-learning